KHorn tweeter physical alignment

[USNRET]

So I just minutes ago purchase a set of K77s from a forum member. I will be using these to experiment with driver alignment. I will have a buddy fashion some walnut mounts to sit on top of the Khorns.

 

I am thinking of having the wooden mounts adjustable in tilt. How should they be installed into the mounts? Flange on the front? Recessed on front so the flange is even with wood? From the back?

 

 

[wvu80]

You might consider mounting them in a baffle to get them up off the top hat, reducing any kind of reflection issues.

[USNRET]

I was thinking of making a complete enclosure (with some sort of tilt adjustment). I have a left over 11 x 17 inch piece of Bob's black grill cloth that can be used. So as to not disturb the horn exit how should it be mounted in the baffle? Front? Back? Back with square cut?

[wvu80]

25 minutes ago, USNRET said:

I was thinking of making a complete enclosure (with some sort of tilt adjustment).

I know you said you had a buddy who could make an enclosure, but what do you think of a small CNC flat pack from PE?  For twenty-five bucks the pair it's hard to go wrong.  You could add a couple of banana plug posts and it would look real clean.

 

Depending on how what orientation you use for the K-77 this might be a stand alone solution for getting the tweeters up in the air.

 

 

Here are larger choices:

http://www.parts-express.com/brand/denovo-audio-inc/521

 

 

[Schu]

I would just do it in a rough manner (for initial testing) front mounted even though those were rear mounted originally. You can always make a permanent enclosure later. 

 

I would love to hear about the audible impressionist of the physical alignment.

 

One more thing... should they be raised/elevated?

[Don Richard]

What are you going to do about the mid-woofer alignment?

[USNRET]

1 hour ago, Don Richard said:

What are you going to do about the mid-woofer alignment?

I will be turning the horns around facing in various directions so that the drivers align vertically.
Just something to play with Don. The experiment is cheap enough.

[CECAA850]

12 hours ago, USNRET said:

Recessed on front so the flange is even with wood?

I would think this ^^^^^

[Chris A]

12 hours ago, USNRET said:

How should they be installed into the mounts? Flange on the front? Recessed on front so the flange is even with wood? From the back?

Front mount.  You don't have to worry about reflections off the top hat--just put down a thin layer of material like cloth or some sort of quilted fabric.  Frequencies above 4500 Hz don't reflect very well...

[jorjen]

23 hours ago, USNRET said:

I will be turning the horns around facing in various directions so that the drivers align vertically.
Just something to play with Don. The experiment is cheap enough.

Remember you may have to deal with phase issues that may arise depending on how far you actually move the tweeter. Only so much can be done running passives. I have run into the same situation with physically aligning my La Scalas. I have noticed a benefit however.

 

This came up in a conversation with Roy a long time ago, can't remember all the details.

[jorjen]

Not trying to jack your thread Mike, just thought you may like to see these and know that others are working on this too..

 

I can move the tweeters on the threaded rods forward or backward. If the tweeters have to move forward a lot the threaded rods are removed and replaced facing the front of the cabinet. Gives me about 10" of total adjustment by just moving the nuts and/or rods. Still experimenting but think it is pretty darn close by ear.

 

Volti VX crossovers and 120 tweeters with A-55 mids from Mr. Crites.

[jorjen]

All that sits atop these Mike.

 

The upper assembly is separate and on cone feet so I can slide it back a lot(slides right over the woofer terminals)to play with aligning the mid horn to the bass cabinet. The crossovers can be placed on a small shelf I will be attaching to the back of the cabinet if the upper assembly has to be moved that far back to achieve alignment which may very well be the case. I have not begun to play with this yet.

 

These things sound really, really good!

[Chris A]

Note that a wavelength at 4500 Hz at room temperature is ~3 inches.  A quarter wavelength at 4500 Hz is 3/4", which is about the limit of misalignment that is tolerable before timbre shifts and soundstage issues begin to become very audible while playing a sine wave at the center crossover frequency.  I find that 1/4 wavelength is actually a little bit too much: I usually aim for ±1/10 wavelength alignment instead.

 

If you're instead crossing at 6000 Hz, then a wavelength is 2.3", and 1/4 wavelength is about 5/8".  Klipsch Heritage loudspeakers typically cross over somewhere in this frequency range.  The K-55 driver is usually not electrically low-passed (except in the newest versions of Heritage crossovers), so you're at the mercy of the driver itself and the condition of its diaphragm to determine the crossover frequency.  The high pass filter on the tweeter (IIRC) is gentle slope, so the interference band between the tweeter and midrange is pretty wide - something like an octave centered on the center crossover frequency between the two drivers.

 

The crossover filters themselves introduce further phase shifts: generally it's 90 degrees of phase lag on the lower frequency driver for every order of the crossover filter.  For a fourth order crossover, that's one full wavelength (2.2"), for second order, that's 1.1 inches (assuming 6000 Hz crossover frequency). 

 

Understand that at these frequencies, either measurement using inverted phase on one driver, or simply playing a sine wave at the center crossover frequency and listening to the timbre shift is usually required in order to achieve these kind of alignment accuracies.  I've found that either method works quite well, as well as using excess group delay measurements using something like REW.

 

Chris

[mikebse2a3]

1 hour ago, Chris A said:

Note that a wavelength at 4500 Hz at room temperature is ~3 inches.  A quarter wavelength at 4500 Hz is 3/4", which is about the limit of misalignment that is tolerable before timbre shifts and soundstage issues begin to become very audible while playing a sine wave at the center crossover frequency.  I find that 1/4 wavelength is actually a little bit too much: I usually aim for ±1/10 wavelength alignment instead.

 

If you're instead crossing at 6000 Hz, then a wavelength is 2.3", and 1/4 wavelength is about 5/8".  Klipsch Heritage loudspeakers typically cross over somewhere in this frequency range.  The K-55 driver is usually not electrically low-passed (except in the newest versions of Heritage crossovers), so you're at the mercy of the driver itself and the condition of its diaphragm to determine the crossover frequency.  The high pass filter on the tweeter (IIRC) is gentle slope, so the interference band between the tweeter and midrange is pretty wide - something like an octave centered on the center crossover frequency between the two drivers.

 

The crossover filters themselves introduce further phase shifts: generally it's 90 degrees of phase lag on the lower frequency driver for every order of the crossover filter.  For a fourth order crossover, that's one full wavelength (2.2"), for second order, that's 1.1 inches (assuming 6000 Hz crossover frequency). 

 

Understand that at these frequencies, either measurement using inverted phase on one driver, or simply playing a sine wave at the center crossover frequency and listening to the timbre shift is usually required in order to achieve these kind of alignment accuracies.  I've found that either method works quite well, as well as using excess group delay measurements using something like REW.

 

Chris

 

 

It should be pointed out that this alignment precision is based on one point in space of the listener's ears vertical height...!   Persons of different heights might fall outside of the alignment window and require a different alignment. 

 

Design mounting for the minimum vertical distance between the 2 sources and longer distances form the sources is your friend in that the further the listener's distance is from the 2 sources the vertical alignment window for the listeners is widened.

 

 

miketn

[Chris A]

The fact that the two driver centerlines are more than 1/4 wavelength apart will create lobing. Time alignment will push the central lobe into alignment while listening on-axis: